The United States Government has a royalty-free license in this invention for its own use and the right in limited circumstances to require the patent owner to license others as provided for by the terms of the Small Business Innovation and Research Contract N00014-94-C-0236.
1. Field of the Invention
This invention relates to electronic circuits and particularly to printed circuits on conventional polymer-based substrates.
2. Description of the Related Art
Printed circuits are universally used to interconnect and "package" active and passive electronic components to perform a myriad of useful functions. These circuits can be of the rigid type which consist of laminated "boards" usually of glass-reinforced epoxy, but also of paper-phenolic and other constructions. There are also flexible circuits based on unreinforced polymer films such as Kapton polyamide manufactured by DuPont. In both cases the circuits are produced by a combination of etching and deposition technologies based on the use of photoresists which are exposed to light and developed to permit a pattern of conductors to be etched and/or plated on the substrate.
These conventional processes are described, for example, in "Printed Circuit Board Basics", M. Flatt, Miller Freeman, San Francisco, 1992. The conventional technology requires many steps of plating, both electroless and electrolytic, and etching to produce a finished board. At each of these stages there is opportunity for error, and hazardous wastes are generated in many of the processes. The principal wastes can be the result of etching away unwanted conductor material to leave behind the desired conductor pattern, and from spent solutions from electrolytic plating of copper, tin and lead, as well as spent electroless plating solutions. The latter are the most difficult to handle because of the presence of formaldehyde reducing agent, a known carcinogen, as well as the toxic heavy metals.
There is a major emphasis on speed of response in the circuit board industry, and technology which can reduce the time to produce prototype boards and facilitate modifications with minimum tooling change, as well as reduce in-process inventory and manufacturing time would be advantageous. In particular, a low temperature analog of "thick film" technology which is used to print circuit conductors onto ceramic substrates which are fired at high temperatures is desirable.
The ideal solution to this problem would be an ink which could be applied only where a conductor is needed by a simple printing technology to any type of printed circuit substrate material, and which could be cured at a temperature which would not damage the substrate. The conductor should have an electrical conductivity greater than half that of the bulk metal and should be strongly adherent to the polymer substrate after curing. This ink would allow the production of circuit boards at high production rates with the generation of no hazardous waste by a simple two-step print and heat process.
The primary problem in the industry is the requirement for high electrical conductivity with a low enough curing temperature to be compatible with the polymer-based circuit boards. Conventional solutions to the problem provide low temperature conductive epoxies with poor electrical conductivity and high temperature thick film inks with good electrical conductivity which can only be used on ceramic substrates. These small, expensive and specialized substrates can withstand the required thick film ink firing temperatures of more than 650.degree. C. and usually above 850.degree. C. An ink which could duplicate this performance on polymer-based substrates at approximately 250.degree. to 350.degree. C. would permit application of this technology broadly in the 20 billion dollar worldwide rigid circuit board industry and the one billion dollar worldwide flexible circuit industry.
"Thick film" technology, as described by R. W. Vest in "Electronic Ceramics", R. Breckenridge, ed., 1991, is routinely practiced to produce hybrid circuits on ceramic substrates. The conductor patterns are created by silk screening or stencil printing thick film pastes or inks onto ceramic substrates and firing them at temperatures of 850.degree. to 1100.degree. C. to reduce the metal-containing inks to metal. An example of such inks are silver-palladium compositions described by Wang, Dougherty, Huebner and Pepin, J. Am. Ceram. Soc. 77(12), 3051-72 (1994). Typically thick film inks contain metal powders, an inorganic glass binder and a vehicle consisting of a polymer binder and a solvent. The vehicle provides the correct consistency for screen printing and consists typically of a polymer such as ethyl cellulose, hydrogenated rosin or polyacrylics dissolved in a low volatility solvent. Common solvents are terpineol, dibutyl carbitol and various glycol ethers and esters. The inks are applied to ceramic substrates by screen printing, dried to drive off the solvent and heat treated, usually in a belt furnace, to decompose the polymer binder and fuse the metal and the inorganic glass binder. The glass phase provides the bond to the substrate which is usually alumina, and the metal provides the electrical conductivity. Typically the conductors have a striated cross section with layers of glass alternating with layers of metal. The glass tends to concentrate at the ceramic interface and the metal at the air interface. The conductivity is typically one half to one quarter that of the bulk metal.
A number of thick film compositions contain surfactants to improve screenability and stability of the metal powder dispersions. Often these surfactants are metallo-organic compounds such as soaps of carboxylic acids. These are convenient in that they decompose at relatively low temperature to deposit the metal or its oxide which can perform a useful function in the fired conductor.
For example, U.S. Pat. Nos. 5,071,826 ('826 patent) issued on Dec. 10, 1991 and 5,338,507 ('507 patent) issued on Aug. 16, 1994 to J. T. Anderson, V. K. Nagesdh and R. C. Ruby and assigned to Hewlett Packard Co. of Palo Alto, Calif. describe the addition of silver neodecanoate to superconducting oxide mixtures to increase the critical current for multicrystalline ceramic superconductor materials. The neodecanoate is decomposed to the metal at 300.degree. C. to coat the superconducting grains with silver. The coated grains are then sintered and oxidized at 600.degree.-800.degree. C. to produce an oxide superconductor of enhanced strength and critical current.
The addition of titanate to thick film conductors by decomposition of an organo- metallic titanate is described by K. M. Nair in U.S. Pat. No. 4,381,945, assigned to E. I. DuPont de Nemours and Co. Wilmington, Del.
U.S. Pat. No. 4,599,277 relates to a process for adding organometallic compounds to thick film inks to increase the densification temperature of the metal to match that of the ceramic substrate at 850.degree.-950.degree. C.
Other conventional thick film paste compositions containing silver flake, glass frit and silver resinates, which are carboxylic acid soaps, as well as surfactants such as Triton X 100, are described in U.S. Pat. Nos. 5,075,262, and 5,183,784. The metal resinate was found to promote adhesion and minimize cracks and voids in bonding semiconductor dies to a ceramic substrate at 350.degree.-450.degree. C.
U.S. Pat. Nos. 4,130,671, assigned to the U.S. Department of Energy, discloses a similar composition of glass frit and silver resinate which was decomposed at low temperature to provide silver-coated glass particles similar to the superconductor described in the '826 and '507 patents. The particles are applied to a substrate either before or after decomposition of the resinate and fired in an oxidizing atmosphere at 500.degree. to 700.degree. C. to provide a conductor of metal-coated glass particles.
Still other conventional thick film compositions of glass and metal powders in an organic vehicle but without the resinate are described in U.S. Pat. Nos. 5,250,229 and 5,378,408 assigned to DuPont, de Nemours and Co. The above described conventional thick film compositions have the shortcoming of requiring high temperatures, i.e., greater than about 450.degree. C. to bind the composite to the substrate.
To create a low temperature analog of the thick film process, it will be necessary to find a new mechanism to obtain adhesion and cohesion of the deposited metal which can operate at temperatures below 450.degree. C., which is the extreme upper temperature limit that polymers can tolerate. The use of inorganic glass powder binders which are universally used in conventional thick film inks is not possible in this application because none of them melt at a low enough temperature, and the glass will not bond to the metal or to the polymer substrates.
Some approaches to creating electrically conductive inks for application to polymer substrates have been described. The most common one is the creation of conductive epoxies or conductive inks by incorporating metal powder, usually silver powder, in an organic matrix. This is a major industry with products available from Ablestik, AIT, Hokurika, M-Tech, Thermoset, Epoxy Technology and Ferro, among others. These materials can be printed on circuit boards, and they have good adhesion. An example of the application of this technology was described in an article by K. Dreyfack in Electronics 52(17), 2E-4E, 1979, on Societie des Produits Industrielles ITT who silk screened silver and graphite-based conductors of this type onto rigid and flexible circuits. The problem with this approach is that the inks conduct by random contacts between powder grains in the organic matrix, and the conductivity is poor. Typical values of the resistivity, which is the reciprocal of conductivity, are 40 to 60 microhm cm, compared to bulk silver at 1.59 microhm cm and high temperature thick film conductors at 3-6 microhm cm. A typical resin-bonded copper powder conductor is described in Japanese Patent Application 52-68507, June, 1977.
U.S. Pat. No. 4,775,439, assigned to Amoco Corp., Chicago, Ill., describes a more elaborate approach with the same results. A metal powder and binder are applied to a substrate and dried. The trace is then covered by a polymer film which is adhesively laminated to the substrate to hold the conductor in place. However, this patent does not address the problem of obtaining electrical conductivity comparable to bulk metal.
Near bulk conductivity has been achieved at low temperature by decomposing metallo-organic compounds on various substrates. They can be applied by ink jet printing as described by R. W. Vest, E. P. Tweedell and R. C. Buchanan, Int. J. (Vest et al.) of Hybrid Microelectronics 6, 261-267, 1983. Vest et al have investigated so-called MOD (Metallo-Organic Decomposition) technology over many years. The most relevant aspect of this research was reviewed in "Liquid Ink Jet Printing with MOD Inks for Hybrid Microcircuits" Teng, K. F., and Vest, R. W., IEEE Transactions on Components, Hybrids and Manufacturing Technology, 12(4), 545-549, 1987. The authors described their work on printing silver and gold conductors as well as dielectrics and resistors. MOD compounds are pure synthetic metallo-organic compounds which decompose cleanly at low temperature to precipitate the metal as the metallic element or the oxide, depending on the metal and the atmosphere. The noble metals, silver, gold and the platinum group decompose to metal films in air. The organic moiety is bonded to the metal through a heteroatom providing a weak link that provides for easy decomposition at low temperature. An oxygen bond, as in carboxylic acid-metal soaps, has been found to be satisfactory, as have amine bonds for gold and platinum.
Vest et al. investigated metallization of ceramic substrates and silicon by ink jet printing of xylene solutions of soaps such as silver neodecanoate and gold amine 2-ethylhexanoate. Images of satisfactory resolution (0.003 inches or 75 microns) were obtained, but the conductivity was low because of the extremely small thickness of the layers after decomposition which was less than a micron.
Preliminary experiments by the inventor of the present application on epoxy-glass circuit boards with silver neodecanoate solutions demonstrated that well-bonded conductors could be produced on polymer substrates. Again, the difficulty was that they were very thin and had inadequate conductivity. It was found that the addition of more MOD compound resulted in wider traces but not thicker ones. The MOD compound melts before decomposing and spreads over the surface uncontrollably. Since melting provides for a well-consolidated metal deposit after decomposition, which is desirable, and since some MOD compounds are actually liquids at room temperature, this is an unavoidable problem. A possible solution to this problem is to build up the thickness by printing many layers, which Vest et al found suitable for metallizing silicon solar cells, but this detracts from production of circuits in a single pass which is the desired result.
Similar materials and techniques have been used to apply thin film metallization and seed coatings which are then built up with solder or electroplating. U.S. Pat. No. 4,650,108, issued on Mar. 17, 1987, to B. D. Gallegher and assigned to the National Aeronautics and Space Administration, Washington, DC; U.S. Pat. No. 4,808,274 issued on Feb. 28, 1989, to P. H. Nguyen and assigned to Engelhard Corp. Menlo Park, N.J.; U.S. Pat. No. 5,059,242 issued on Oct. 22, 1991 to M. G. Firmstone and A. Lindley and U.S. Pat. No. 5,173,330 issued on Dec. 22, 1992, to T. Asano, S. Mizuguchi and T. Isikawa and assigned to Matsushita Electric Co, Ltd. Kadoma, Japan, are examples. However, the above described thin films alone cannot provide adequate conductivity.
A creative attempt to circumvent the resistivity problem was described in U.S. Pat. No. 4,487,811 issued on Dec. 11, 1984, to C. W. Eichelberger and assigned to General Electric Co. Schenectady, N.Y. This patent describes augmenting the conductivity by a replacement reaction of metal in the deposit by a more noble metal in solution, for example the replacement of iron by copper. In the process of doing this, the contact between particles is improved by the greater volume of the replacement metal and its greater intrinsic conductivity. A resistivity of 7.5 microhm cm was achieved, substantially better than silver-loaded epoxies, but short of the performance of thick film inks. The replacement reaction solved yet another problem of polymer inks in that the material was solderable, which conductive epoxy formulations in general are not.
Another approach to solderability was described in U.S. Pat. No. 4,548,879 issued on Oct. 22, 1985 to F. St. John and W. Martin and assigned to Rohm and Haas Co. Philadelphia, Pa. Nickel powder was coated with saturated monocarboxylic acid with ten or more carbon atoms. The coated powder was mixed with novolac epoxy resins in a butyl carbitol acetate vehicle and silk screened onto an epoxy-glass board. After curing at 165.degree. C., the conductive trace could be solder-coated by fluxing and dipping into molten solder, while a trace made with uncoated nickel powder could not be soldered. No improvement in electrical conductivity was described with this process.
U.S. Pat. No. 4,186,244 ('244 patent) issued Jan. 29, 1980, and U.S. Pat. No. 4,463,030 ('030 patent) issued Jul. 31, 1984 to R. J. Deffeyes, and H. W. Armstrong and assigned to Graham Magnetics, Inc. North Richland Hills, Tex. relates to silver powder compositions having a low film-forming temperature. The silver powder was formed by decomposing dry silver oxalate in the presence of a long chain carboxylic acid, either saturated (stearic acid, palmitic acid) or unsaturated (oleic acid, linoleic acid). The acid reacted with the metal powder as it was formed to provide a protective coating on the surface and to limit the particles to sub-micron size. The particles were washed to remove excess acid and blended with an equal weight of a conventional thick film vehicle consisting of ethyl cellulose polymer binder and pine oil solvent.
The resulting ink was coated on a ceramic or polyamide substrate and heated to 250.degree. C. in air for 30-90 seconds to convert the coated powder to a silver conductor with a conductivity of one ohm per square. The coating is said to be solderable without flux, which is believable if residual acid is acting as a flux. It is stated to be resistant to leaching in a bath of molten solder, which is unexpected, based on the well known solubility of silver in solder. The explanation may lie in the quoted conductivity, which is a thousand-fold less than that required for practical circuits.
Example I of the '244 patent indicates that the resistance of the resulting material drops markedly during the "visible fusing" period, from about 6 ohms per square to one ohm per square. The "visible fusing" period is stated to occur when the substrate is heated to 250.degree. C. in air for 90 seconds during which the conductor pattern changes to a silvery-white color. No thickness for the deposit is stated, which would allow a calculation of the resistivity of the silver. Most thick films are of the order of 25 microns (0.001 inch) thick, and this is generally used as the standard for comparison. At 25 microns thickness one ohm per square corresponds to a resistivity of 2500 microhm cm compared to the resistivity of bulk silver which is 1.59 microhm cm. This suggests a very poorly consolidated deposit with entrained nonconducting material. Even a one micron thick film corresponds to a resistivity of 100 microhm cm. A resistance of one ohm per square is far too high for practical circuitry which typically has traces with lengths of many hundreds of squares.
The unconsolidated nature of the deposit may be due in turn to the very tenacious stearate coating on the silver particles, which is an object of the '244 and '030 patents. It is well known that stearates and similar materials, which are commonly added to silver and other powders to prevent agglomeration, are very difficult to remove, even at temperatures of 625.degree. C. and above. If not removed, they will inhibit sintering and increase resistivity. This subject was discussed in "Effect of Particle Size Distribution in Thick Film Conductors", R. W. Vest, Proceedings of the Flat Plate Solar Array Project Research Forum on Photovoltaic Metallization Systems Nov. 15, 1983 DOE/JPL-1012-92.
A somewhat similar silver flake material was described in U.S. Pat. No. 4,859,241. The flake was prepared by milling silver powder with silver stearate surfactant in an organic solvent to produce silver stearate- coated silver flakes providing a glass-filled ink composition of superior stability. This is a common method of preparing stable powders and flakes of silver.
None of the materials or mixtures described above accomplish the goal of providing an ink which can be cured to a well-bonded, well-consolidated metallic conductor with an electrical conductivity comparable to conventional thick film inks but with a curing temperature below approximately 350.degree. C., which is required for compatibility with conventional polymer-based circuit board substrates. None of these materials has made it possible to impact the circuit board industry with new technology for rapid production by a simple process with no hazardous waste production. A new approach to provide this low temperature capability is needed.